TW399364B - Method and apparatus for wideband predistortion linearization - Google Patents

Abstract

A Predistortion Signal is generated which compensates for third order and higher order intermodulation distortion over a wideband. The predistortion signal is applied to an input radio frequency signal, for example a multi-tone radio frequency signal. The predistortion signal is a low order polynomial having adjustable coefficients. The predistortion signal can compensate for intermodulation products produced a nonlinear amplifier, and the polynomial coefficients can be adjusted based on the difference between the amplifier output and the input radio frequency signal.

Description

V. Description of the invention (1) [Field of the invention] _ The present invention relates to a method and device for distortion compensation. More specifically, it relates to a method and an apparatus for linearizing a wide-band predistortion. [Background of the Invention] Radio frequency (RF) signals often include envelope variations caused by, for example, amplitude modulation or combining two or more RF carriers or RF audio. If these amplitude-modulated or multi-tone RF signals are amplified by, for example, a non-linear amplifier, they can cause intermodulation distortion (IMD). IMD can cause unwanted interference at frequencies other than the audio of a multi-tone RF signal. This interference = often occurs at frequencies close to the audio ’and is therefore difficult to filter out. Therefore, some form of linearization is needed to suppress the IMD caused by the non-linear amplifier. There must be a compromise between distortion performance and efficiency in amplifier design. Linear amplifiers operating under "level A" have very little distortion but are not efficient, while non-linear amplifiers operating under "level C" have very reasonable efficiency but introduce severe distortion. Although both efficiency and distortion are important considerations in amplifier design, the importance of efficiency increases at high power levels. Non-linear amplifiers are widely used because of their efficiency, but the problem of distortion must be addressed. Several amplifier linearization techniques are known to reduce distortion caused by non-linear amplifiers. Traditional amplifier linearization techniques can be broadly classified as feedback, feedforward, or predistortion. Feedback is a known linearization technique that has been widely used. For example, U.S. Patent No. 2, 102, 671 to Black revealed an early negative reaction

V. Description of the invention (2) Feed linearization technology to reduce distortion on radio frequency. H. A. Rosen and A. T.

Owens' "Power Amplifier Linearity Studies for SSB Transmissions", IEEE Transactions on Communication Systems, pp. 15 0-1 59, June 1 964 discloses a feedback linearization technique to reduce RF distortion. A more recent feedback technique is Cartes ia η feedback using negative feedback of quadrature modulation in the baseband. An example of such feedback is disclosed in " Stability Analysis of Cartesian Feedback Linearization for Amplifiers with Weak Non 1 inearities " by MA Briffa and M. Faulkner, IEE Proceedings on Communications, Vol. 143, No. 4, pp_ 212-218, Aug. 1 9 9 6. Other modern feedback linearization techniques include polar feedback, discussed in US Patent No. 5, 〇23, 937 granted to opas, and intermediate frequency (IF) feedback, as disclosed in k. G. Voyce and JH Me Candle ss Power Amplifier Linearization

Using IF Feedback ", IEEE MTT-S Digest, pp. 863-866, 1989. The problem with these feedback linearization techniques is that system latency often limits the available linearization bandwidth. Feedback techniques are therefore often limited to narrow-band systems, such as single-carrier linear modulation designs. Another disadvantage of these feedback linearization techniques is their potential instability. RuFe is another known linearization technique that has been successfully applied to RF. A typical feedforward RF power amplifier is disclosed in U.S. Patent No. 5,157,346 granted to Powel et al. According to this technique, the amplified output is compared with the input signal in a first comparison loop to obtain an error signal. Error signal

5. Description of the invention on page 6 (3) The amplifier is amplified and added to the output at the phase of the second correction loop degree, whereby the technology provides a disadvantageous error amplifier of linearization before widening linearization frequency, and therefore The efficiency of the feedforward technology is low, although the former level A amplifier is more efficient. Another known linearization technique Linearization is achieved by making the conversion from pre-loss to a linear function by distorting the pre-amplifier. In the middle, the distortion of the original output is offset by 180 to the original output distortion. This wide range has excellent IMD suppression. The reason is that it usually requires the use of a line to operate under level A. This will degrade the amplifier and still have the same performance than pre-distortion. According to this technique, the predistortion technique that distorts the input signal with the distortion input from the amplifier to the amplifier output can be applied to the true RF frequency or baseband, that is, before modulation with the RF carrier. When applied to {^ frequency, the pre-distortion technique exhibits linear performance over a wide frequency band. However, because the predistortion function is more complex and therefore more difficult to implement on higher-order distortions, predistortion techniques are usually reduced to a maximum of only the third-order IMD component. Cuber Predistortion Linearizer for Relay by T. Nojima and T. Konno

Equipment in 800 MHz Band Land Mobile Telephone Systems ", IEEE Transactions on Vehicular Technology, Vol. VT-34, No. 4, pp. 169-177, Nov. 1 98 5 and US Patent No. 4 granted to No. ima Nos. 943, 783 disclose typical cubic predistorters that reduce IMD up to the third order. Fig. 1A illustrates a conventional cube predistorter 1 as disclosed by No. ima and Konno. The input RF signal received by input terminal 2 is divided into 4 by the power divider. 5. Description of the invention (4) '-are two signals with substantially the same amplitude. One of the divided signals is applied to a linear signal path including a variable delay line 9. The other divided signal is applied to a non-linear signal path including a third-order function generator 6, a variable phase adjuster 7, and a variable attenuator 8. The third-order function generator 6 generates a third-order predistortion signal according to the input RF signal, and outputs the predistortion signal at the terminal 12. The variable phase adjuster 7 adjusts the phase of the predistortion signal, and the variable attenuator 8 adjusts the amplitude of the predistortion signal. The amplitude and phase adjusted predistortion signal is combined in the combiner 5 with the linear signal provided by the delay line 9. The combined signal is transmitted at the terminal 10 to the RF power amplifier (M) 13. In this way, the third-order MD component caused by RF pa 丄 3 will be eliminated from the amplified signal, so that it is linearized with pA i 3 'If the delay line 9 compensates for the delay caused by the predistortion signal, a wide frequency can be achieved Linearization of the band. _, Figure 1B shows the spectrum of the RF signal composed of fl and cardiac audio, which can be applied to Pre-loss 1. Figure 1c illustrates the output spectrum of the RF PA 13. As shown by = -C, the output spectrum contains the basic components of ^ and d as represented by a solid line and the third-order iMD distortion octave produced by 2 ^-; ^ and pa 13. The output spectrum also contains a third-order predistortion component with a frequency of 2fi_f2 & 2f2_fi, which is injected by the pre-distortion 77 1, and is represented by a dashed line. As shown in the figure, the injected third-order predistortion component has the same amplitude but opposite phase to the third-order component of PA13. Therefore, the third-order predistortion component is prepared to cancel the third-order IMD component. ~ One of the problems with this method is that RF power amplifiers rarely have only the third-order IMD component and do not produce higher-order MD components. Generally, 5. Invention Description (5) Higher-order IMD distortion components are not valued, and only when the third-order IMD components are suppressed can they really appear. To solve this problem, S. P. Stapleton and J. K. Cavers' " A New Technique for the Adaptation of

Linearizing Predistorters ", Preceedi ngs of the IEEE Vehicular Technology Conference, pp. 753-758, May 1991 discloses predistorters that compensate for third and higher order IMD components. Figure 2 shows this improved predistorter, which basically consists of three main blocks: the quadrature gain phase adjuster (QGPA) 1 4, the baseband polynomial predistorter circuit (PreD) 1 5 and the controller 1 6 .

As shown in Figure 2, the input RF signal is sent to qGPA 14 and PreD 15 in the input terminal 丨 7. The PreD 15 circuit uses the detector 22 to detect the envelope of the input heart signal and processes the envelope of f by the non-linear function generators Fi (x) 23 and f2 (x) 24. The function generator Fi (x) 23 and dagger (also) 24 will generate in-phase and positive father (I & q) signals, and multiplex with the incoming signal in QGPA 14 to form the third and fifth-order pre- Distortion component. The use of the complex multiplexing process network makes the frequency modulation (AM / AM) and phase modulation (AM / PM) distortion in Cartesian form. In order to achieve composite multiplexing, the input rf # is first cut. For two paths, one enters the multiplexer 20. The signals on the two paths are between " with phase π being complete, but the input to the multiplexer 21 is phase shifted by the private phaser 20 to be 90 degrees from the signal input to the multiplexer ⑺. The two polynomials generated by 2 4 are in the function generator Fi (x) 23 & F2 (x) 5. Explanation of the invention (6) ~~ The function coefficients are provided by the microprocessor (βρ) 2 5 in the controller 16 . The microprocessor 25 will adjust the coefficients based on the in-phase and quadrature feedback signal stages obtained from the rotation of the linearized rf PA (not shown). I & ^ The feedback signal is band-pass filtered in filters 28 and 29, and is measured by separating imd from the required number. This is only possible when the desired # signal being amplified is a single carrier signal, because at this time it is known that the components will lie on the frequency bands on both sides of the single carrier modulation. Detectors 2 6 and 2 7 will determine the number of IMD bands with dense passbands so that the microprocessor 25 adjusts the coefficients of the predistortion function to minimize 'distortion' and thereby will appear in the electrical components of the imd component of the RF PA output. Flat minimize. Although this technique compensates for third and higher order I components, it is only suitable for single carrier applications. In a multi-carrier system, the position of the carrier and its IMD may not necessarily be obtained by bandpass filtering. This makes the technology unsuitable for wideband applications, which usually involve multiple audio input signals. The wideband nature of these applications is derived from a combination of individual band evaluations " is 5 tigers apart from a specific band plan (spectrum). Digital signal processing (DSP) can be applied to more detailed predistortion: for example, US Pat. No. 5, 049, 832 to Cavers discloses an adaptive linearization technique for f = DSP. The problem with the secret P is the linear :: frequency: it is quite limited by the sampling frequency of the DSP and the required digital / analog conversion-: so ’a system using a DSP is generally not suitable for wideband applications. Band library technology cannot compensate for the wideband IMD with multiple audio input signals. In addition, these traditional techniques cannot distinguish the signal of the bee value and the average level.

V. Description of the invention (7): ~~~~~-When the δ audio cascade is in a multi-audio system, the peak value of the constructive dry audio RF signal is much larger than its average level. Non-linear multi-tone power method? The driver must be economical and reasonable in power efficiency, so there is no way to accommodate this peak. The result is that some peaks in the signal will be cut off when the amplifier saturation limit exceeds the frequency. 9 In RF in the predistorter shown in Figures 1A and 2, there is no type 1 diagram to distinguish the operation under or above Γ. As a result, the predistortion is caused to saturate the RF input signal of the RF PA in the hanging condition. The application of a free and 'pre-distortion #' will not have an output amplitude. ^ According to the huge compensation input of PA, it will not limit its output of the second vibration. Of course, the traditional predistorter in terms of phase, these phase corrections are relative to the loss: the efficiency is greatly reduced, so it will cause predistortion performance in these ridge values.耵 Wide-band predistortion technology for chirp signal, = 1 MD component of Guang Er @ order. There is also a need for a pre-distortion technique that is effective for the peaks of RF input kbl. [Summary of the invention] Zhejiang. The object of the present invention is to provide a technique for compensating high-order 1 MD components. A further goal is to provide an average input signal with low IMD compensation. At the same time, Tian Leliang prevents the IMD compensation of the peak input signal from decreasing. According to a specific embodiment of the present invention, a predistortion signal is generated to compensate

5. Description of the invention (8) Compensate the third-order and higher-order 1 components over a wide frequency band, and the pre-distortion signal will be applied to an input RF signal, such as a multi-tone RF signal. The predistortion function is a low-order polynomial with adjustable coefficients. The detected envelope is clipped to a near hyperbolic tangent waveform and its proportion is adjusted. The limiting wave pattern approximates a hyperbolic tangent (t a n h) function to prevent the detected envelope from exceeding a certain value, and thus prevent predistortion compensation that is relatively inexact when a large peak appears in the detected envelope. According to a specific embodiment 'the predistortion signal compensates for the IMD components produced by the non-linear amplifier. The polynomial coefficients are adjusted based on the difference between the amplifier output and the input RF signal. [Brief description of the drawings] With the accompanying drawings and the following detailed description of specific embodiments of the present invention, these and other objects, features, and advantages of the present invention can be more prominent. In the figure:, FIG. 1A, traditional third Block diagram of the first-order predistorter; Spectrum B is the spectrum of the dual audio input RF signal of the predistorter applied to Figure A. Figure C is the RF PA amplifier wheel of the second-order predistorter in Figure 1-. The block diagram shown in Figure 1 is an analog predistortion system of a specific embodiment. Figure 2 is a traditional fifth-order predistorter. Figure 3 is a block diagram according to the present invention. Road details Ξ = Ϊ ΐ According to the present invention-specific implementation Pre-distortion

Page 12 -------- V. Description of the invention (9) Figures 5A to 5C show the behavior of predistorted RF signals according to the example of CTS-specific embodiments, Figures 6A to 6 in the frequency domain Six c is the behavior of the pre-distorted RF signal according to the example of this example. A specific embodiment, in the time domain, FIG. 7A is the method according to the present invention '~' -β; and-the eight-body embodiment compensates the distortion Exemplary Figure 7B is an exemplary method according to the present invention. —A specific embodiment generates a pre-distortion signal. [Detailed description of the preferred embodiment] In the following description, the pull-in * technique, etc., is for illustration rather than control = such as a specific circuit, circuit element, or solution. Skilled in this technology: To provide a thorough understanding of the basic characteristics of the present invention, the present invention can be implemented without departing from its spirit and known methods, dream storage, and special features. In other cases, 2, and the circuit The detailed descriptions are omitted to avoid the details of @f from obstructing the description of the present invention. For the sake of the letter, the third and higher-order IMD components will be obtained by putting the pre-mistake directly on the input 1 ^ signal. It is compensated. According to the specific example of the present invention, the analog processing element that allows wideband operation will generate predistortion ^ 10 / Figure 2 is an example predistortion system according to a specific embodiment of the present invention. The system includes Analog predistortion circuit, and the quadrature gain phase adjuster (QGPA) 36 that sends the predistortion signal to the audio pA 13 as much as possible.

PreD will be based on how many audio RF input signals are received at the terminal 30.

Page 13

Properly distorted signals are generated internally. q is divided into two phases by the turn-on light coupler 33. The input received at terminal 4 and terminal 30 is shown as a composite signal with a time-dependent 5-amplitude branch. (The thick-line meter path is fed back to the delay element 42. It is also divided for observation.)-/ Segmenter 34 ^ Splitter 34 feeds the branch indicated by ^^^ into a * thing 〇 遽 x and cuts it to have Two-two Γ two f of the same amplitude provide PreD37, and the other one feeds two-degree phase division. The 5-handed 90-degree phase splitter divides the signal χ into two branches and multiplies: one branch by 0 degrees, and The other branch is multiplied by 90 degrees, so a composite signal is obtained.

PreD 37 again generates a pre-distorted signal p based on the signal. The 90-degree divider 35 causes the QGPj 36 to phase the composite predistortion signal p from the prej) 37 and the signal f · again, thereby adjusting the order and phase of the signal X. The adjusted signal is a signal r ′ which is output from QGPA 36 to rf PA 13. The operation of the predistortion system according to the gain-increasing form can be described mathematically as follows: Γ = χΡ (1) where Γ is the predistorted RF signal, X is the RF input, and P is the predistortion signal generated by PreD 37 (Or dynamic composite gain signal), that is, P = Pi + jpq. The predistortion signal p can be generated by PreD 37 in Cartesian form as follows:

Pi = I x I 2C2i + | X | C1; + Coi (2a)

Pq = lx I 2C2q + | x | Clq + C〇q (2b)

Here I x I is the number of stages of the signal x, and C2i, Ch, CGi, C2q, CiQ, and k represent the coefficients of the predistortion signal, which can be adjusted by the controller 40 D

5. Description of the invention (11) The QGPA 36 may be implemented as, for example, two multipliers and one adder in order to multiply each item in the predistortion signal with the incoming signal χ. The expandable plural form of formula one is rewritten as follows: r = x (lxl 2C2 + | x | C1 + C〇) (3) where C〇 = C〇i + jC ^ (4 a) ci ~ CH + jc] g ( 4b) C2 = C2i + K2q (4c) Refer to Figure 3 again. The controller 40 will adjust the pre-loss by minimizing the difference between the input q and the adjusted RF PA 13 output. Light-coupler 4? And pressure reducer Seoul: The second middle school adjusts the ratio, and the output delay time after the adjustment of the ratio is basically equivalent to 4; true = = element 42 43 uses the number * 1 / Gd to adjust the coupling The delay of the pressure reducer II increases the torque, the benefit of the coupling 47, etc., so that the adjustment of the rear wheel output signal = the gain of the two Γ delayed wheel input signals. The adjusted wheel output: No. = the input signal passes through the light coupling 45 is closed to the detector 41, crying 4. The error is mentioned: The difference between the late input signals is sent to the control c, Γ, and the difference is bluffed 6. The controller 4 will adjust the predistortion The signal coefficients Cu, Ershi 21, and k are used to minimize the error signal e. This can be used to account for changes in amplifier characteristics caused by ®, 庚 庚 ί 庚 ίίΪ. The controller produces 2 = coefficients CQi and, and are added by 38 and 39 Most of the difference between the static and the amplifier input and output round out the pre mind controller 4〇

Page 15 V. Description of the invention (12) It can be implemented with, for example, a microprocessor. As can be seen from the pre-distortion pass number account of Formula 3, that is, Γ-Π, Γ nr η increase profit, there is no PreD 37, that is, Cu-0, ClQ = 0, C2i = 0, and (^ = 〇 = > The gain of QGPA 36 is dominated by the setting of the complex coefficients c-1 2, m, Λη 07 η ± λ number L0 ~ C〇i + COq. Therefore, without PreD 37, QGPA 36 °: To adjust the complex number of RF PA 13. The RF input benefits for the application will change with the RF input level :: The fixed number of RF PA 1 3 will increase, and this fixed adjustment will lead to; It is nonlinear-ψ # ^ „D has an input at an RF level-the business trip is zero. The introduction of preD 37 can dynamically change the 2 flat function, and therefore: the second complex f gain depends on the input voltage. ^, eD n 〇17 U This effective linearization of RF PA 13. Adding Cl allows the pre-D 3 7 to adjust the complex number of QGPA 36 based on the ratio of λ and λ k, batch 2 ^, and number.加入. Adding makes the lucky number ^ 4 —p, square of the number. Yueying can be changed in response to the input stage " Figure ::: Detailed block diagram of the predistortion circuit of the offense. According to the specific The predistortion circuit of the second corresponds to FIG. 3M. However, the month is not the limit: and the predistortion circuit according to the present invention requires a higher-order IMD compensation in the system. The input RF signal RF 1 (corresponding to Λ & 2 @ 二 kX) is applied to the envelope after the test including the mixer 48 and the limiter 49; the envelope of the input rf signal will be detected. 50% effective inspection; and amplifier (VSA) 5 °. VSA through the control of the arrow package of terminal c3 to limit, the amplitude of the limiting level used can be close to: 2 voltage control from the outside. According to Specific embodiment 'Limited and tanh (tanh) function can prevent the envelope from exceeding

P.16 5. Description of the invention (13) This prevents the PreD 37 from being tested.

A rather inaccurate pre-correction is made at the peak. A frequency-limiting signal with a large I 50 is applied to the variable gain amplifier vga 5 丨 to adjust the ratio of the frequency-limiting signal, and can make the 37 circuit close or actively suppress in response to the control calendar above the terminal. The output after squaring the V-ratio scale represents the pre-processing packet of the rf input = 2 is squared by 1x1 in the squarer 52 to obtain | χ | 2. In order to produce the functions of equations 2a and 2b, 丨 x | and 丨 χ 丨 2 are applied to four linear summator circuits Γ54, 55, and 56. These multipliers will add the signals multiplied by 1 i 'i and k α by additions 57 and 59 respectively, and store them in buffers 58 and 60 to obtain two ⑥ out of R j and PRQ. Think of the output that represents the higher-order terms of formula 2 & and Devices 38 and ^ are added to PRI and PRQ respectively to obtain Pi and dagger. ... The above is described in terms of gain; the pre-distortion system can also be extended by the input-to-output conversion function of Equation 3 as follows:

丨 χ 丨 2 C2 + X | X xC0 (5) A formula 5 shows that the predistortion signal r contains the first order xC0 that compensates for the first order, and the third order that compensates for the third-order IMD | 2 (: 2, And an additional term for compensating for many IMD components exceeding the third-order IMD component χ | χ | (: ι,

V. Description of the invention (14) '—'— -___________ Effective linearization performance of the South-level IMD component. Refer to the line of the frequency of the true number r after excluding coefficients. For a better understanding of the pre-miss. From Figure 58 and Figure 5c, we can see that =, the first, and the first: there is a first-order predistortion for this, and the χ 丨 χ 丨 2 terms are derived from the X term to improve distortion. The second-order pre-division of the second-order range term is that for each term, such as the X5 order term, the example is more order term, and the example of the circuit is multi-tone. The pre-distortion of the order 1MD component, the xixi term is dynamic. From Figs. 6A to 6C, the predistortion signals in the time domain 1 see that the amplitude of these items changes more quickly, and the rise and fall are faster. Similarly, when χ | χ 丨 is compared with other high-order, X, etc., when the amplitude of χ is quite large (H), the other high-amplitude rises faster than the amplitude of the XI χ 丨 term, so it will be more than the electrical formula = The frequency limit was reached early. When x is quite small (< 丨), other high amplitude drops are also faster than the amplitude of the X I x 丨 term, so it will approach the noise floor faster than the electronic one. Therefore, the χ | χ I term can simplify the pre-distortion electronic embodiment, especially when processing signals with inherently high dynamic range. Fig. A is an example of compensating for distortion according to a specific embodiment of the present invention. The method 1111 starts at step 7 0 0, where an input multi-audio r F signal is received: in step 720, the predistortion signal is produce. At step 74, the pre-distortion signal is applied to the input multi-tone RF signal. After applying the predistortion signal, the input multi-audio RF signal will be amplified, for example, and the IMD generated by the amplifier will be compensated by the predistortion signal. As shown in Fig. 7a, as long as a round RF signal is received, the predistortion method is repeated again and again.

Page 18 V. Description of Invention (15) 'I ~~-'. ^ 6 describes an exemplary method for generating a predistortion signal according to a specific embodiment of the present invention. The method starts at step 722, where the envelope of the input multi-audio RF wave is detected. The detected envelope is frequency-limited in step 724, and the integer ratio is 5 weeks in 7 2 6. Secondly, in step 7 2, the controller 4 adjusts the system according to, for example, errors detected between the amplifier turn-in and the amplifier output. Finally, at step 73, the controller 40 multiplies the coefficient adjusted by 1 using the detection envelope to calculate a polynomial predistortion signal. According to the present invention, it is possible to improve the average IMD performance over a wide frequency band and reduce incorrect predistortion at the peak input level. In an experimental test at a center frequency of 1500 Mjz and an average round-out power of 25 WaUs (w), an average improvement in scallop (dB) was obtained without destroying the peak 1 MD performance at least 10 Λ. The pre-distortion technique according to the present invention can be applied to almost any carrier frequency with only minor modifications. Implementation T T has been described in a specific embodiment, but is described herein. : Description of headings, not for limitation. < Listed as above; to compensate for the distortion caused by the non-linear amplifier 1 ^ _ The invention can be applied to the compensation of IMD from any source. In addition, although the "input signal" is described as a multi-tone RM ancient solution, the present invention can also be applied to a single-tone input signal. Because == 口 .IS fixed form to implement. The scope of the present invention = yes, the "The month" is not limited by the foregoing description. The meaning of this case and the equivalent scope are within the scope of the present invention. Another change should be wished by Ma Han.

Page 19

Claims (1)

β? 1 · 20? g ^ > VI. Patent application scope: a kind of predistortion circuit that compensates distortion over a wide frequency band; and intermodulation and distortion components of intermodulation 2 The predistortion signal is applied to the input radio 2. According to the first item of the scope of patent application, t 諕 * road. Is more. Audio RF signals. "Let the input RF signal 3. According to the device of the first scope of the patent application, the circuit of No. is constituted by a quadrature gain phase adjuster: The pre-distortion signal is applied 4 'According to the device of the first scope of the patent application, low-order Polynomial. 〒3 Pre-distortion signal is 5. According to the scope of the patent application, the pre-distortion operation is generated according to the following formula: The pre-distortion circuit is ... 丨 X 丨 2 .... uICi; xC. Number, f, RF signal , X means input radio frequency signal 6. Control of distortion signal produced by fit 利范 77 fit1. Two: Patent | & the device around the first item, # 中 此 调 = ii sex amplifier generated, and predistortion The signal will be applied to the input RF signal first in a non-linear manner. Line amplifier 8. The device according to item 7 of the scope of patent application, wherein the controller q ϊϋΐί and the difference between the input RF signal adjust the predistortion signal. 9. According to the device in the patent application No. 1 item, the distortion circuit contains an envelope detector that detects the envelope of the input radio frequency signal, and will detect the envelope according to = 1__ Page 20 Generate a predistortion signal. 10. The device according to the ninth item of the scope of the patent application, wherein the predistortion circuit includes a variable saturation amplifier to limit the detected envelope frequency to an approximately hyperbolic sine waveform. 1 1. According to the tenth scope of the patent application The device of claim 1, wherein the predistortion circuit includes a variable gain amplifier that adjusts the ratio of the frequency-limited envelope. 1 2. A method for compensating distortion over a wide frequency band, including the following steps: generating a predistortion signal; and the predistortion signal; Applied to the input RF signal, where the pre-distortion signal will compensate the third-order and higher-order intermodulation distortion components. 1 3. The method according to item 12 of the patent application scope, wherein the input RF signal is a multi-audio RF signal. 1 4. The method according to the twelfth item in the scope of the patent application, wherein the predistortion signal is a low-order polynomial. 1 5. The method according to the twelfth item in the scope of the patent application, wherein the predistortion signal is generated according to the following formula: r = x | x | 2C2 + x | x | Cj + xC0 where r is the pre-distorted input RF signal, x is the input RF signal, and C2, q, and Co represent possible 1 The method according to item 12 of the scope of patent application, further comprising the step of adjusting the predistortion signal. 1 7. The method according to item 12 of the scope of patent application, wherein the intermodulation distortion component is It is generated by a non-linear amplifier, and the predistortion signal is applied to the input RF signal before the non-linear amplifier is amplified. Page 21 6. Scope of patent application 1 8. The method according to item 17 of the scope of patent application, wherein the predistortion signal is adjusted according to the difference between the amplifier output and the input RF signal. 19. The method according to item 12 of the scope of patent application, wherein the step of generating a predistortion signal further comprises the following steps: detecting an input RF signal envelope; and generating a predistortion signal based on the detected envelope. 20. The method according to item 19 of the scope of patent application, wherein the step of generating a pre-distortion signal further comprises the step of limiting the detected envelope frequency to an approximately hyperbolic sinusoidal waveform. 2 1. The method according to the twentieth item of the patent application, wherein the step of generating a pre-distortion signal further includes a step of adjusting a detected envelope ratio. Page 22